Amplifier- Saturation

[dr.power]

Hi guys,

When designing an amplifier out of op-amps or power amplifier chips, How do you notice and guarantee the input signal will not causes the amplifier to go to saturation?
I know how to notice that for "sine" waves but do not have any clue for the audio or speech.
I hope my question make sense to you. Most of tests while designing an amplifier are held by injecting a SINE wave to the amplifier, at such a saturation you can just consider the
Vp-p and the gain of the amplifier to guess and avoid the saturation statement. You can even do not calculate that but just connect the input to a sine wave generator and monitor the output by an scope and change the pot of the signal generator to see when the amplifier goes to the saturation and then find out the maximum input amplitude you can inject to the amplifier. But as we all know an amplifier is designed to work by complicated waves like speech and the audio. Then what to do?

Thanks a lot

 

[ronsimpson]

Some high end audio amps have a dual voltage comparitor on the output.
Example: +&-25 volt supplies, the amplifier can drive +/-20 volts, the voltage comparitors are set to +18 and -80 volts. Any time the output is beyond 18 volts a LED lights for 0.5 seconds.

In broadcast it is critical that too much volume is not fed into the transmitter. Limiters are used to "turn down" the volume and turn up the volume to keep the peaks from being too much and to push up the low levels.

Which topic do you want to talk about? Something different?

 

[dr.power]

I am talking about when You are about to design an amplifier.

Example: +&-25 volt supplies, the amplifier can drive +/-20 volts, the voltage comparitors are set to +18 and -80 volts. Any time the output is beyond 18 volts a LED lights for 0.5 seconds.

You meant "-18"?
You mean some thing like sound meters, as I told I have faced such confusing problem when I want to design an amplifier.
By the way I have seen some final product amplifier telling:
""Max Input for Max Output .......1.0Vp-p, each channel @ 1kHz""
or
""Input sensitivity......... 6 dBm (1.55 V) or -14 dBm (150 mV)""
Saying those input parameters which are for whiles the input is a pure sine is helpful at all?

 

[ronsimpson]

Turn up the volume until it sounds bad then turn it down a little.

In the case of college students...Turn up the volume until it sounds bad then turn it up more.

Amplifiers are designed around sign wave power. What standard is there?

Audio/speech is very different. A live concert has very large dynamic range. That is “fixed” when recorded and/or broadcast. Someone will yell about my choice of words (fixed). During the recording process the dynamic range is changed so it sounds good on a tape/CD or radio. Each type of audio has a very different ratio of peak voltage (power) to average power and each type has a different sensitive to distortion.

I don’t think I told you much except there is no answer.

 

[KeepItSimpleStupid]

The AMP is designed to work with a specific input signal and has a particular frequency response where it won't clip. That's the common name for saturation for an amplifier. The signal "clips" when the positive or negative portion of the signal is flat. When it's flat, DC is passed.

With an audio power amplifier, the presence of DC is detected at the amp by looking at the effective charge on a capacitor exceeding some value. This protection circuit disconnects the speakers. The AMP has a high pass filter at it's input, so it won't amplify DC. Clipping indicators can be designed into amplifiers as well. Fuses are not normally used, but can be. The reason why they aren't used is because the consumer can replace it with a higher value and cause more problems.

Certain amplifiers have what they call dynamic headroom where the amplifier rails can change with signal content, so a short-acting peak won't clip.

 

[audioguru]

An amplifier is supposed to be designed with plenty of power so it never clips. But usually an amplifier is designed to save money so it does not have enough power then it clips when the output signal is trying to be louder than is possible. Clipping sounds awful because high levels of distortion are produced.

 

[alec_t]

Automatic gain control (AGC) can be applied to an amplifier so that the output can't exceed a certain level. But AGC by its nature introduces distortion. Audibly, this distortion may be less objectionable than clipping of the output.

 

[Warpspeed]

Hi guys,

When designing an amplifier out of op-amps or power amplifier chips, How do you notice and guarantee the input signal will not causes the amplifier to go to saturation?

There are only two places an amplifier can clip, either right at the input, or at the power output stage.

The input has a common mode input voltage range that must not be exceeded, and the output stage has both a peak output voltage swing and very often a peak current limit that must not be exceeded either.

These voltages and currents are clearly shown on the data sheet for the particular chip.

 

[audioguru]

Play a tone, vocals or music through an amplifier that drives a speaker or headphones. Slowly turn up the volume.
When the amplifier begins to clip then there will begin to be severe distortion.
The distortion become worse as you turn up the volume because the amplifier output is trying to be beyond its max output and its output waveform is a square wave.

 

[Nigel Goodwin]

Automatic gain control (AGC) can be applied to an amplifier so that the output can't exceed a certain level. But AGC by its nature introduces distortion. Audibly, this distortion may be less objectionable than clipping of the output.

You don't apply AGC, you apply limiting - these are VERY different things.

Limiting doesn't affect the audio in any way, UNTIL it's triggered, and then it simply turns the volume down.

You don't use it on domestic amplifiers (there's no need), but generally it's used on PA amplifiers, which will usually also display a light if the limiting is reached.

 

[alec_t]

You don't apply AGC, you apply limiting - these are VERY different things.

I am aware of the difference. Limiting may be the conventional approach, but AGC is also possible. By 'turning down the wick' at the amplifier front-end gradually as the signal level at the input increases (e.g. by logarithmic compression), clipping and hard-limiting in output stages can be avoided. I mentioned AGC as a possible approach, not knowing precisely what level of 'Fi' the OP is aiming for.

 

[Nigel Goodwin]

I am aware of the difference. Limiting may be the conventional approach, but AGC is also possible. By 'turning down the wick' at the amplifier front-end gradually as the signal level at the input increases (e.g. by logarithmic compression), clipping and hard-limiting in output stages can be avoided.

You don't 'hard limit', only soft - AGC affects quality (or at least 'changes' it), limiting doesn't (until the limiting point is reached).

But this is more a PA thing anyway - presumably the OP is attempting to make an SSB transceiver (although he's never mentioned it) - the audio stages are only basic anyway, quality isn't any concern, just recognisable speech.

Many, many, years ago Plessey used to make a special IC for this exact purpose, along with a completel range for amateur radio use.

 

[audioguru]

Radio and TV stations use AGC for audio instead of limiting. Then loud sounds cause the gain to be reduced which will be too low for low level sounds that flollow until the AGC slowly increases the gain. Low level background noise by itself results in the AGC increasing the gain until thje louder background noise is irritating.
Some AGC circuits are too slow which results in a blast of very loud severe distortion until the AGC circuit reduces the gain.

 

[crutschow]

Automatic gain control (AGC) can be applied to an amplifier so that the output can't exceed a certain level. But AGC by its nature introduces distortion. Audibly, this distortion may be less objectionable than clipping of the output.

A good AGC circuit doesn't introduce distortion (a least not of the harmonic variety) it just reduces the gain. (Unless, of course, you consider a reduction in the music's dynamic range, distortion).

 

[alec_t]

A good AGC circuit doesn't introduce distortion (a least not of the harmonic variety) it just reduces the gain. (Unless, of course, you consider a reduction in the music's dynamic range, distortion).

I suspect at least some AGC circuits in low-end consumer equipment don't fall into the 'good' category . As for whether they introduce harmonic distortion that presumably depends on the speed of response. In theory a fast-acting AGC might change the gain slightly (very slightly!) within one cycle of a low-frequency signal, and that would alter the signal waveform. In practice, of course, the AGC time constant should be chosen to make such alteration negligible.

 

[ronsimpson]

Many audio amplifiers do not clip well. Example 1 below shows the amplifier coming out of clipping slow and then over shoots. In example 2 there is asymmetrical clipping and in the positive direction phase reversal. Example 3 shows what some tube or valve amplifiers do.
(Sorry about the bad CAD drawings)
One advantage of a limiter is predictable clipping that can match a valve amplifier. The limiter is set just below the power amplifier's clipping level.

 

[audioguru]

If your amplifier clips then either you have its volume control turned up too high or it doesn't have enough power for the volume you want.

 

[ronsimpson]

If you are playing canned music (CD, tape, FM or God for bid AM) then the level is controlled for you. Some one in a studio with limiters and compressors fixed the peak points. If you are live then the peak points can be many tens of db above what you want.
First thing a person does on stage is pound on a mic (with a fist) to see if it is live. "Is this thing on?" really really loud! When I have 16 wired mics, a bank of wireless and a 16 channel recorder running and I can't find who is thumping on their mic......it is easy to distort a dual 500 watt amp.
I run good peak limiters/compressors ahead of the rack of power amplifiers. The limiters push the peaks down and the average up. The audience want loud! and a little soft clipping is fine! "Turn it up" so I slam harder into the limiters, the average goes up and no one knows I saved money on the amps/speakers.

 

[KeepItSimpleStupid]

Those guitar enthusiasts WANT that tube clipping. A simple DC clip just doesn't do it for them.

 

[alec_t]

Reminder: the OP wants to know how, at the design stage, to guarantee the amp won't go into saturation (i.e. clip). So, in summary, incorporating either compression (AGC) or soft-limiting into the designed amp is the answer.